The health and long-term survival of the growing population of childhood cancer survivors is threatened by late sequelae of therapy: cardiovascular complications, obesity, osteoporosis, and new malignancies. This application proposes a re-examination of a methodologically sophisticated clinical trial data set that has the potential to generate new hypotheses about the antecedents of health-risk behavior and health-promotion behavior in adolescent survivors of childhood cancer. The analysis will use a tested, broad-based, middle range theory of health behavior to identify potential new determinants of these behaviors in previously unexplored and/or reconfigured variables from the parent clinical trial. Unlike the parent trial, the proposed study will examine the health risk behaviors and health protective behaviors as 10 separate behaviors and as two multidimensional constructs, rather than as a single, unidimensional construct. This economical investigative approach maximizes the probability of identifying new targets for interventions aimed at reducing survivors' risk of late sequelae. Specifically, this study will: (1) define the relationships between selected demographic, social, psychological, and resource variables and the health risk behaviors and health promotion behaviors of adolescent survivors of childhood cancer; (2) specify two separate structural equation models (health risk behavior and health promotion behavior) in adolescent survivors of childhood cancer by using the Interaction Model of Client Health behavior (IMCHB); and (3) compare the explanatory power of the cognitive processing theory (Health Belief Model, which directed the parent clinical trial) with that of the IMCHB, which will direct the proposed study. This study will afford an established investigator (PI) an opportunity to enter a newly targeted high-priority area (health promotion and quality of life in people with chronic illness) with a self-contained research project that uses novel methods of analysis (reconceptualization of measures, logistic and multiple regression, item response theory, and structural equation model development). This study has the potential to make a major conceptual contribution to childhood cancer research and to provide the conceptual basis for prospective clinical trials of health promotion interventions for survivors of childhood cancer.